[dpdk-dev] [PATCH v6] vhost: support virtqueue interrupt/notification suppression

[Chen, Junjie J]

Hi

> -----Original Message-----
> From: Yuanhan Liu [mailto:yliu at fridaylinux.org]
> Sent: Monday, January 8, 2018 10:07 PM
> To: Chen, Junjie J <junjie.j.chen at intel.com>
> Cc: Wang, Xiao W <xiao.w.wang at intel.com>; maxime.coquelin at redhat.com;
> Bie, Tiwei <tiwei.bie at intel.com>; dev at dpdk.org; Yao, Lei A
> <lei.a.yao at intel.com>
> Subject: Re: [PATCH v6] vhost: support virtqueue interrupt/notification
> suppression
> 
> On Tue, Dec 26, 2017 at 12:43:10PM -0500, Junjie Chen wrote:
> > The driver can suppress interrupt when VIRTIO_F_EVENT_IDX feature bit
> > is negotiated. The driver set vring flags to 0, and MAY use used_event
> > in available ring to advise device interrupt util reach an index
> > specified by used_event. The device ignore the lower bit of vring
> > flags, and send an interrupt when index reach used_event.
> >
> > The device can suppress notification in a manner analogous to the ways
> > driver suppress interrupt. The device manipulates flags or avail_event
> > in the used ring in the same way the driver manipulates flags or
> > used_event in available ring.
> >
> > This patch is to enable this feature in vhost.
> >
> > Signed-off-by: Junjie Chen <junjie.j.chen at intel.com>
> 
> You need put "---" before the change log. Otherwise, it will be tracked in the
> commit log.

Will update this.
> 
> > +#define vhost_used_event(vr) \
> > +	(*(volatile uint16_t*)&(vr)->avail->ring[(vr)->size])
> > +
> > +static __rte_always_inline void
> > +vhost_notify(struct virtio_net *dev,  struct vhost_virtqueue *vq) {
> > +	/* Don't notify guest if we don't reach index specified by guest. */
> > +	if (dev->features & (1ULL << VIRTIO_RING_F_EVENT_IDX)) {
> > +		uint16_t old = vq->signalled_used;
> > +		uint16_t new = vq->last_used_idx;
> > +
> > +		LOG_DEBUG(VHOST_DATA, "%s: used_event_idx=%d, old=%d,
> new=%d\n",
> > +			__func__,
> > +			vhost_used_event(vq),
> > +			old, new);
> > +		if (vring_need_event(vhost_used_event(vq), new, old)
> 
> It's a bit weird that you use one from the standard linux header file
> (vring_need_event), while you define you own one (vhost_used_event).
> Note that the system header file also has "vring_used_event()" defined.
The vring_used_event is defined and used for virtio in kernel, kernel defines a vhost_used_event in vhost.c for vhost, so I just use a separated macro for vhost end.

I'd like to define both vhost_need_event and vhost_used_event in vhost.h to remove potential build issue in old linux distribution and also to keep consistent. Is that OK for you?

> 
> Besides that, I have few more comments (and some requirements):
> 
> - It'd be much better if there is a Tested-by tag. Expeclitly,
>   I'm asking a test with Linux kernel virtio-net driver in guest.
Sure.
> 
> - I also hope you could have done a build test on some old distributions.
>   AFAIK, the two macros (vring_need_event and vring_used_event) come
>   from kernel 3.0 (or above). Any kernel older than that would fail
>   the build.
> 
> - I'd be great if you could make a new one based on top of my latest
>   tree: I have just applied a patchset that should conflict with this
>   one.

Sure, will do this.
> 
> 	--yliu